The V4L2 fwnode data structure doesn't get freed on unbind, which leads to
a memleak.

Fixes: e43ccb0a045f ("media: i2c: Add support for the OV5648 image sensor")
Signed-off-by: Rafael Mendonca <rafaelmendsr@gmail.com>
---
 drivers/media/i2c/ov5648.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/media/i2c/ov5648.c b/drivers/media/i2c/ov5648.c
index dfcd33e9ee13..95850f06112b 100644
--- a/drivers/media/i2c/ov5648.c
+++ b/drivers/media/i2c/ov5648.c
@@ -2598,6 +2598,8 @@ static int ov5648_remove(struct i2c_client *client)
 	mutex_destroy(&sensor->mutex);
 	media_entity_cleanup(&subdev->entity);
 
+	v4l2_fwnode_endpoint_free(&sensor->endpoint);
+
 	return 0;
 }
 
-- 
2.34.1

Hi Rafael,

On Sun 18 Sep 22, 23:32, Rafael Mendonca wrote:
> The V4L2 fwnode data structure doesn't get freed on unbind, which leads to
> a memleak.

Thanks for this patch, good catch!

I agree with Jacopo, you shouldn't add a newline before the call.
Feel free to send a v2 with that fixed and include my:

Reviewed-by: Paul Kocialkowski <paul.kocialkowski@bootlin.com>

Cheers,

Paul

> Fixes: e43ccb0a045f ("media: i2c: Add support for the OV5648 image sensor")
> Signed-off-by: Rafael Mendonca <rafaelmendsr@gmail.com>
> ---
>  drivers/media/i2c/ov5648.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/drivers/media/i2c/ov5648.c b/drivers/media/i2c/ov5648.c
> index dfcd33e9ee13..95850f06112b 100644
> --- a/drivers/media/i2c/ov5648.c
> +++ b/drivers/media/i2c/ov5648.c
> @@ -2598,6 +2598,8 @@ static int ov5648_remove(struct i2c_client *client)
>  	mutex_destroy(&sensor->mutex);
>  	media_entity_cleanup(&subdev->entity);
>  
> +	v4l2_fwnode_endpoint_free(&sensor->endpoint);
> +
>  	return 0;
>  }
>  
> -- 
> 2.34.1
> 

-- 
Paul Kocialkowski, Bootlin
Embedded Linux and kernel engineering
https://bootlin.com

On Tue, Sep 20, 2022 at 03:25:20PM +0200, Paul Kocialkowski wrote:
> Hi Rafael,
> 
> On Sun 18 Sep 22, 23:32, Rafael Mendonca wrote:
> > The V4L2 fwnode data structure doesn't get freed on unbind, which leads to
> > a memleak.
> 
> Thanks for this patch, good catch!
> 
> I agree with Jacopo, you shouldn't add a newline before the call.

Thanks for the review, I'll send a v2. As the code for the ov5648 is
similar to the ov8865, and in the 8865 there is this blank line before
the v4l2_fwnode_endpoint_free() call, I thought it might be good to keep
the same style here, that's why this empty line ended up appearing here.

> Feel free to send a v2 with that fixed and include my:
> 
> Reviewed-by: Paul Kocialkowski <paul.kocialkowski@bootlin.com>
> 
> Cheers,
> 
> Paul
> 
> > Fixes: e43ccb0a045f ("media: i2c: Add support for the OV5648 image sensor")
> > Signed-off-by: Rafael Mendonca <rafaelmendsr@gmail.com>
> > ---
> >  drivers/media/i2c/ov5648.c | 2 ++
> >  1 file changed, 2 insertions(+)
> > 
> > diff --git a/drivers/media/i2c/ov5648.c b/drivers/media/i2c/ov5648.c
> > index dfcd33e9ee13..95850f06112b 100644
> > --- a/drivers/media/i2c/ov5648.c
> > +++ b/drivers/media/i2c/ov5648.c
> > @@ -2598,6 +2598,8 @@ static int ov5648_remove(struct i2c_client *client)
> >  	mutex_destroy(&sensor->mutex);
> >  	media_entity_cleanup(&subdev->entity);
> >  
> > +	v4l2_fwnode_endpoint_free(&sensor->endpoint);
> > +
> >  	return 0;
> >  }
> >  
> > -- 
> > 2.34.1
> > 
> 
> -- 
> Paul Kocialkowski, Bootlin
> Embedded Linux and kernel engineering
> https://bootlin.com

Hi Rafael,

On Tue 20 Sep 22, 10:58, Rafael Mendonca wrote:
> On Tue, Sep 20, 2022 at 03:25:20PM +0200, Paul Kocialkowski wrote:
> > Hi Rafael,
> > 
> > On Sun 18 Sep 22, 23:32, Rafael Mendonca wrote:
> > > The V4L2 fwnode data structure doesn't get freed on unbind, which leads to
> > > a memleak.
> > 
> > Thanks for this patch, good catch!
> > 
> > I agree with Jacopo, you shouldn't add a newline before the call.
> 
> Thanks for the review, I'll send a v2. As the code for the ov5648 is
> similar to the ov8865, and in the 8865 there is this blank line before
> the v4l2_fwnode_endpoint_free() call, I thought it might be good to keep
> the same style here, that's why this empty line ended up appearing here.

Understood! I'd be tempted to make a cosmetic fix for 8865 but will probably
wait to group it with functional changes.

Cheers,

Paul

> > Feel free to send a v2 with that fixed and include my:
> > 
> > Reviewed-by: Paul Kocialkowski <paul.kocialkowski@bootlin.com>
> > 
> > Cheers,
> > 
> > Paul
> > 
> > > Fixes: e43ccb0a045f ("media: i2c: Add support for the OV5648 image sensor")
> > > Signed-off-by: Rafael Mendonca <rafaelmendsr@gmail.com>
> > > ---
> > >  drivers/media/i2c/ov5648.c | 2 ++
> > >  1 file changed, 2 insertions(+)
> > > 
> > > diff --git a/drivers/media/i2c/ov5648.c b/drivers/media/i2c/ov5648.c
> > > index dfcd33e9ee13..95850f06112b 100644
> > > --- a/drivers/media/i2c/ov5648.c
> > > +++ b/drivers/media/i2c/ov5648.c
> > > @@ -2598,6 +2598,8 @@ static int ov5648_remove(struct i2c_client *client)
> > >  	mutex_destroy(&sensor->mutex);
> > >  	media_entity_cleanup(&subdev->entity);
> > >  
> > > +	v4l2_fwnode_endpoint_free(&sensor->endpoint);
> > > +
> > >  	return 0;
> > >  }
> > >  
> > > -- 
> > > 2.34.1
> > > 
> > 
> > -- 
> > Paul Kocialkowski, Bootlin
> > Embedded Linux and kernel engineering
> > https://bootlin.com
> 
> 

-- 
Paul Kocialkowski, Bootlin
Embedded Linux and kernel engineering
https://bootlin.com

Hi Rafael

On Sun, Sep 18, 2022 at 11:32:46PM -0300, Rafael Mendonca wrote:
> The V4L2 fwnode data structure doesn't get freed on unbind, which leads to
> a memleak.
>
> Fixes: e43ccb0a045f ("media: i2c: Add support for the OV5648 image sensor")
> Signed-off-by: Rafael Mendonca <rafaelmendsr@gmail.com>
> ---
>  drivers/media/i2c/ov5648.c | 2 ++
>  1 file changed, 2 insertions(+)
>
> diff --git a/drivers/media/i2c/ov5648.c b/drivers/media/i2c/ov5648.c
> index dfcd33e9ee13..95850f06112b 100644
> --- a/drivers/media/i2c/ov5648.c
> +++ b/drivers/media/i2c/ov5648.c
> @@ -2598,6 +2598,8 @@ static int ov5648_remove(struct i2c_client *client)
>  	mutex_destroy(&sensor->mutex);
>  	media_entity_cleanup(&subdev->entity);
>
> +	v4l2_fwnode_endpoint_free(&sensor->endpoint);
> +

Nit: all the other function calls are in a single block, this one
doesn't deserve an empty line before it.

Nit apart, the patch is a good fix.


>  	return 0;
>  }
>
> --
> 2.34.1
>

Hi Rafael,

On Sun, Sep 18, 2022 at 11:32:46PM -0300, Rafael Mendonca wrote:
> The V4L2 fwnode data structure doesn't get freed on unbind, which leads to
> a memleak.
> 
> Fixes: e43ccb0a045f ("media: i2c: Add support for the OV5648 image sensor")
> Signed-off-by: Rafael Mendonca <rafaelmendsr@gmail.com>
> ---
>  drivers/media/i2c/ov5648.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/drivers/media/i2c/ov5648.c b/drivers/media/i2c/ov5648.c
> index dfcd33e9ee13..95850f06112b 100644
> --- a/drivers/media/i2c/ov5648.c
> +++ b/drivers/media/i2c/ov5648.c
> @@ -2598,6 +2598,8 @@ static int ov5648_remove(struct i2c_client *client)
>  	mutex_destroy(&sensor->mutex);
>  	media_entity_cleanup(&subdev->entity);
>  
> +	v4l2_fwnode_endpoint_free(&sensor->endpoint);
> +
>  	return 0;
>  }
>  
> -- 
> 2.34.1
> 

Looks good to me.
Reviewed-by: Tommaso Merciai <tommaso.merciai@amarulasolutions.com>

Regards,
Tommaso

-- 
Tommaso Merciai
Embedded Linux Engineer
tommaso.merciai@amarulasolutions.com
__________________________________

Amarula Solutions SRL
Via Le Canevare 30, 31100 Treviso, Veneto, IT
T. +39 042 243 5310
info@amarulasolutions.com
www.amarulasolutions.com